Public, free and always growing
Learning resources
Hand-picked, free material for teaching yourself robotics. A growing collection, starting with one path from Python to a walking robot.
The starter path
From Python to a walking robot, at about ten hours a week. Post a build log after each step.
- 00Start hereHow the learning resources work, what you need on your desk, and the order of things.1 evening
- 01Programming you'll actually useThe terminal, Git and Python, to the level where tooling stops being the problem.2 to 4 weeks
- 02Just enough mathLinear algebra first, then the calculus and probability that robot learning leans on.3 to 5 weeks
- 03Robotics fundamentalsFrames, kinematics and control, from the two textbooks that taught most of the field.6 to 8 weeks
- 04Your first armThe SO-101 with LeRobot. Teleoperate, record a dataset, train a policy, watch it run.4 to 6 weeks, with an arm
- 05SeeingEnough deep learning and computer vision to understand what your policy is looking at.4 to 6 weeks
- 06Learning to actReinforcement learning, in simulation, with MuJoCo. The engine behind walking robots.6 to 8 weeks
- 07Legs and sim-to-realWhy a policy that walks in simulation falls over on the floor, and how the Open Duck Mini closes the gap.6 to 10 weeks
- 08Wheels, heads and the wider worldMobile bases, desktop robots that talk, and how to show your work so it counts.Open-ended
Found something better? Suggest it in the Discord or by email.